Taphonomic Aspects of Crowned Hawk-Eagle Predation on Monkeys

Journal of Human Evolution 44 (2003) 87–105

William J. Sandersa*, Josh Trapania,b, John C. Mitanic

a Museum of Paleontology, University of Michigan, 1109 Geddes Avenue, Ann Arbor, MI 48109-1079, USA b Department of Geological Sciences, University of Michigan, 2534 C.C. Little Building, 425 E. University, Ann Arbor, MI 48109-1063, USA c Department of Anthropology, University of Michigan, 500 South State Street, Ann Arbor, MI 48109-1382, USA

Received 24 January 2002; accepted 11 November 2002

Abstract

This study provides a taphonomic analysis of prey accumulations of crowned hawk-eagles (Stephanoaetus coronatus) from Ngogo, Kibale National Park, Uganda, collected over 37 months from below nests of two eagle pairs. Crowned hawk-eagles are powerful predators capable of killing animals much larger than themselves, and are significant predators of cercopithecoid monkeys in forest habitats throughout sub-Saharan Africa. At Ngogo, 81% of the individuals in the kill sample are monkeys. Redtail monkeys (Cercopithecus ascanius) are particularly well represented in the sample, making up 66% of monkeys identified to species. Despite an impressive killing apparatus, crowned hawk-eagles are fastidious eaters that inflict far less damage to bone than mammalian predators. Examination of skeletal material from the Ngogo kill sample reveals that crania, hindlimb elements, and scapulae survive predation better than do other bones. Crania of adults are typically complete and accompanied by mandibles, while crania of young individuals are usually dissociated from mandibles and lack basicrania and faces. Long bones are often whole or show minimal damage. Thin bones, such as crania and innominates, are marked by numerous nicks, punctures, and “can-opener” perforations. Scapular blades are heavily raked and shattered. Along with the strong preference for cercopithecoids, these distinct patterns of bone survival and damage indicate the feasibility of recognizing specific taphonomic signatures of large raptors in fossil assemblages. Berger and Clarke (1995) hypothesized that crowned hawk-eagles or similar large raptors were principally responsible for the accumulation of the late Pliocene fossil fauna from Taung, South Africa, including the type infant skull of Australopithecus africanus. The results of our study suggest that the faunal composition and type of damage to the hominid skull and other bone from Taung are consistent with the predatory activities of large raptors. More rigorous assessment of their hypothesis will require sorting the Taung fauna by locality and further detailed analysis of species composition and bone damage and survivability patterns. 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Crowned hawk-eagle; Predation; Taphonomy; Cercopithecoid monkeys; Taung

* Corresponding author. Tel.: +1-734-647-2098 E-mail addresses: [email protected] (W.J. Sanders), [email protected] (J. Trapani), [email protected] (J.C. Mitani).

Introduction Recently, crowned hawk-eagles were implicated as possible accumulators of the faunal assemblage Crowned hawk-eagles (Stephanoaetus corona- from the late Pliocene site of Taung, South Africa tus) are large (weight, 3.4–4.1 kg; length, 81– (Berger and Clarke, 1995), which includes cerco- 92 cm), immensely powerful raptors (Steyn, 1973; pithecoid bones and the infant type skull of Williams and Arlott, 1980; Brown et al., 1982). Australopithecus africanus (Cooke, 1990). Results Armed with deep, well-curved beaks, thick legs, of our analysis have important implications for robust toes, and long, sharp talons that are report- this and other such hypotheses. Along with show- edly stronger than those of any other African ing a strong preference by resident crowned hawk- eagle, they are voracious predators of small-to- eagles for cercopithecoids, skeletal analysis of the medium-sized mammals, and can kill animals as Ngogo kill aggregation reveals distinct patterns of large as bushbuck (ca. 30 kg) (Brown, 1971; Steyn, bone survival and damage. These findings indicate 1973; Daneel, 1979; Brown et al., 1982; Steyn, the feasibility of recognizing specific taphonomic 1983). They are significant predators of cercopithe- signatures of large raptors in fossil assemblages. coid monkeys throughout sub-Saharan Africa (Skorupa, 1989; Struhsaker and Leakey, 1990; Leland and Struhsaker, 1993, Maisels et al., 1993; Sample and methods Hart et al., 1996; Mitani et al., 2001), a fact attested to by the variety and intensity of mecha- The kill sample was collected between July 1996 nisms used by monkeys to deter predation by these and August 1999 from below the nests of two pairs raptors, including loud calls, tightened grouping of of crowned hawk-eagles. Details about nestlings troop members, rapid descent through the canopy, and fledglings associated with each nest are and aggressive defensive behavior (Gautier-Hion reported in Mitani et al. (2001). Searches at the and Tutin, 1988; Cordeiro, 1992; Maisels et al., first nest (Nest A) were conducted approximately 1993). once a week during nestling phases and periods of In this paper we provide a taphonomic inter- eaglet dependence, and at other times bi-weekly. pretation of prey assemblages accumulated by The second nest (Nest B) was discovered in June crowned hawk-eagles living in the Ngogo study 1999 and monitored bi-weekly over three months. area in Kibale National Park, Uganda and col- Searches involved comprehensive, diligent, and lected at their nests over a period of 37 months repetitive surface collecting. Material was not (Mitani et al., 2001). Eight diurnal anthropoid screened, but many small bones and bone frag- species are present at Ngogo and elsewhere in ments were collected. Because of the diligence of Kibale National Park, including a hominoid the searches for prey remains and collecting under (chimpanzees, Pan troglodytes), two colobines the nests, it is likely that the study sample accu- (black and white colobus, Colobus guereza, and red rately reflects the production of bone from the colobus, Piliocolobus badius), and five cerco- nests. pithecines (redtail monkeys, Cercopithecus ascan- The assemblages collected below the nests ius, blue monkeys, C. mitis, l’Hoest’s monkeys, comprised of bone, hair, boluses, eggshell, and C. lhoesti, olive baboons, Papio anubis, and grey- feathers. We collected a total of 387 skeletal speci- cheeked mangabeys, Lophocebus albigena) (Leland mens (NISP-number of individual specimens), in- and Struhsaker, 1987; Struhsaker and Leakey, cluding 345 specimens from Nest A and 42 1990; Mitani et al., 2000, 2001; Mitani and Watts, specimens from Nest B. Specimens include complete 2001). As primates, particularly cercopithecoids, and incomplete bones, as well as articulated sets of compose a large percentage of the mammalian bones. Where possible, the following attributes were biomass in the 12 km2 Ngogo area (Struhsaker, noted for each specimen: (1) identity of bone(s); (2) 1997), it is not surprising that monkeys dominate side; (3) taxon; (4) age-sex class; (5) dimensions; (6) the resident crowned hawk-eagle kill sample association; and (7) damage. Taxonomic assess- (Mitani et al., 2001). ments were made on the basis of pelage, size, and W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 89 morphology of bones and teeth, in comparison relative abundance of a certain element i, Ni is with data from museum collections and published the observed number of a particular element (the sources. For example, colobine femora were sorted MNE value) in the taxon of interest, MNI is the from those of guenons and Lophocebus by differ- minimum number of individuals in the taxon of ences in size, degree of projection of the greater interest, and Ei is the number of skeletal element i trochanter, neck length, and intercondylar breadth expected in a complete skeleton (see Andrews,

(see Nakatsukasa, 1994a,b, 1996). 1990; Lyman, 1994b). Ri values are expressed as a Age classes of adult, subadult, juvenile, and percentage representing bone survivability relative infant + neonate were based on size, degree of to what would be expected if all skeletons of all epiphyseal fusion, and tooth eruption. Postcrania individuals had been preserved. It is important to were assumed to be from adult animals if epiphy- point out, though, that because our MNI values ses were completely fused to diaphyses. In bones of were sometimes summed over different age and sex adult or nearly adult size, but where epiphyses classes, in many cases no Ri value approaches were cleanly missing or incompletely joined, indi- 100%. viduals were assigned to the preadult category. To investigate bone fragmentation, we looked Skulls were considered adult if all molars had at the percentage of each element that was whole erupted or if cranial sutures were closed. Skeletal (unfragmented). For each taxon of interest, we elements were sexed by size and robustness divided the number of whole bones of each (see Gautier-Hion, 1975). The soundness of these element by the NISP to arrive at a value, which we criteria was confirmed by observations in museum express as a percentage. One hundred percent collections of substantial dimorphism in dimen- means that the number of whole elements equals sions and extent of muscle markings, particularly the NISP for that element, and there is no in cercopithecoid long bones and scapulae. In fragmentation. adult skulls, tooth size, particularly of canines, was also considered a good indicator of sex. We calculated the minimum number of ele- Results ments (MNE) for each element from each taxon Prey composition and age-sex class. We follow Badgley (1986) and Lyman (1994a) in defining a skeletal element as a Cercopithecoid monkeys are the predominant single complete bone or tooth. Calculating MNE prey of crowned hawk-eagles at Ngogo (Table 1). may become an involved process, especially with The high percentage of monkeys in the MNI count large samples or highly-fragmented bones (e.g., (68/84 = 81%) closely matches the fraction of mon- Marean et al., 2001). In our assemblage, however, keys in crowned hawk-eagle kill accumulations only a small proportion of specimens were frag- from the Kanyawara area of Kibale studied earlier mented, and there was little question as to whether by Skorupa (1989) and Struhsaker and Leakey fragments represented the same or different (1990), 88% and 84%, respectively. Other crowned elements. We calculated minimum number of hawk-eagle prey from Ngogo includes duikers and individuals (MNI; see Grayson, 1984; Klein and palm civets, as well as single fruit bats, rodents, Cruz-Uribe, 1984) based on our MNE values. and hornbills (Table 1). There are no remains of These values were used to investigate bone chimpanzees in the kill accumulations, and no survivability and bone fragmentation in the assem- attacks by crowned hawk-eagles on chimpanzees blage. We compared bone survivability and frag- were witnessed in over five years of field research at mentation patterns among all individuals in each Ngogo (Mitani et al., 2001). Body mass of the primate taxon and Cephalophus, in primates versus adult mammals in the kill accumulation sample non-primates, and between preadult and adult ranges up to around 11 kg (Table 1); perhaps the primates. To investigate bone survivability, we much larger size of chimpanzee adults (Smith and calculated relative abundances of bones using Jungers, 1997) deters crowned hawk-eagles from the formula Ri =Ni/(MNI X Ei), where Ri is the predation on chimpanzee infants. 90

Table 1 Prey assemblage of crowned hawk-eagles at Ngogo, Kibale National Park, Uganda (July 1996-August 1999). Percent values have been rounded to the nearest 0.1%.

Taxon Adult body mass NISP % total NISP % primate NISP MNI % total MNI % primate MNI Primates Cercopithecus ascaniusa 4.3 kg (m) Nest A-165 47.00 53.10 Nest A-22 36.90 45.60 3.1 kg (f) Nest B-17 Nest B-9 b Cercopithecus lhoesti 6.0–8.0 kg (m) Nest A-10 2.60 2.90 Nest A-3 3.60 4.40 87–105 (2003) 44 Evolution Human of Journal / al. et Sanders W.J. 3.0–4.5 kg (f) Nest B-0 Nest B-0 Piliocolobus badiusb 7.9–10.9 kg (m) Nest A-21 6.70 7.60 Nest A-6 8.30 10.30 6.7 kg (f) Nest B-5 Nest B-1 Colobus guerezab 6.8–11.3 (m) Nest A-3 0.80 0.90 Nest A-3 3.60 4.40 5.4–10.9 (f) Nest B-0 Nest B-0 Lophocebus albigenab 5.7–8.7 (m) Nest A-10 2.60 2.90 Nest A-2 2.40 2.90 3.6–8.2 (f) Nest B-0 Nest B-0 Papio anubisc (young juvenile,z 3.0–6.0 kg) Nest A-2 0.50 0.60 Nest A-1 1.20 1.50 Nest B-0 Nest B-0 Unidentiﬁed cercopithecines Nest A-93 26.10 29.40 Nest A-17 22.60 27.90 Nest B-8 Nest B-2 Unidentiﬁed colobines Nest A-0 2.30 2.60 Nest A-0 2.40 2.90 Nest B-9 Nest B-2 TOTAL: 343* 88.60 68 81.00 Non-primates Cephalophus monticolad 3.5–9.0 kg 17 4.40 3 3.60 Nandinia binotatad 2.0–3.2 kg 4 1.00 2 2.40 Megachiroptera 2 0.50 1 1.20 Rodentia 1 0.30 1 1.20 Mammalia: Indeterminate 13 3.40 5 6.00 Aves: Bucerotidae 2 0.50 1 1.20 Aves: indeterminate 5 1.30 3 3.60 TOTAL: 44* 11.40 16 19.00 TOTAL ASSEMBLAGE: 387* 84 Abbreviations: NISP, number of individual specimens; MNI, minimum number of individuals; m, male; f, female. aBody mass of redtail monkeys at Kibale from Jones and Bush (1988). bBody masses from Delson et al. (2000). cBody mass range for young juvenile baboon based on data for wild baboons in Strum (1991). *includes correction for articulations; counting each element (whether articulated or not) as a separate element would lead to NISP values of 413 for primates, 44 for non-primates, and 457 for the whole assemblage. dBody masses from Kingdon (1997). W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 91

Fig. 1. Minimum number of individuals (MNI) of preadult and adult primate taxa in the Ngogo crowned hawk-eagle kill assemblage.

Of the primate individuals identiﬁable to the kill sample, since there are more females than species, those attributable to Cercopithecus ascan- males in resident Ngogo monkey populations ius (redtail monkeys) are by far the most numerous (Mitani et al., 2001). Adults and preadults are (31/47 = 66%), followed by small numbers of red present in approximately expected proportions in colobus, black and white colobus, grey-cheeked the sample (Mitani et al., 2001). Age classes were mangabeys, l’Hoest’s monkeys, and a juvenile determined for all cercopithecoid prey individuals olive baboon (Fig. 1). There is no evidence of blue (Fig. 1), with adults (31/68 = 46%) and non-adults monkeys in the sample. Thirty-one percent of the (37/68 = 54%) nearly equally represented. Al- cercopithecoid sample is too fragmentary and/or though there are more non-adults than adults in too young to identify to species with certainty, but cercopithecoid populations at Ngogo, these pro- the great majority of these appear to be guenons. portions do not diﬀer from those expected on the In contrast to the strong selectivity of the basis of chance (Mitani et al., 2001). Ngogo crowned hawk-eagles for cercopithecoids, and redtail monkeys in particular, they do not Bone survivability show preferences for animals of a particular sex or age-class. Males comprise 55% (16/29) of the por- For the present crowned hawk-eagle assem- tion of the cercopithecoid sample in which sex blage, there is a clear pattern of bone survivability, could be determined, and are over-represented in dominated by crania and hindlimb bones 92 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105

(Figs. 2–4). The relatively low bone survivability in the sample, a high proportion (4/6) are accom- for most skeletal elements accords with the habit panied by their mandibles. Incomplete crania re- of crowned hawk-eagles to dismember and some- covered under the nests are usually comprised of times cache larger, monkey-sized prey away from calotte pieces and are missing their faces and the nest (Brown, 1971; Jarvis et al., 1980; Brown et maxillae. Presumably, when these are removed by al., 1982; Steyn, 1983; Maisels et al., 1993). This the eagles the mandibles are discarded with them, correlates with a feeding preference at the nest for which would account for the much greater number brains and hindlimbs and a tendency to swallow of partial crania (n = 47) than isolated dentaries smaller elements whole, which are subsequently (n = 4). Mandibles show minimal breakage; it is digested completely (Brown, 1982; Brown et al., significant for comparison with kill assemblages of 1982). Often, the trunk is dismembered and its mammalian predators (e.g., carnivores, chimpan- internal organs are consumed at the kill site zees) that mandibles in the Ngogo sample com- (Brown, 1971). monly retain their rami and angles. Of nine Predictably, smaller or more fragile elements mandibles recovered, just two exhibit damaged such as ribs, vertebrae, tarsals, carpals, phalanges, rami, each only on one side. Crania of adults are and metapodials (metacarpals and metatarsals) are usually intact, but exhibit signs of intense manipu- poorly represented (Figs. 2–4). Forelimb elements lation (Fig. 4). These include talon nicks (small (humerus, radius, ulna) are much less numerous V-shaped holes), punctures (more rounded holes) than hindlimb elements (femur, tibia, fibula). In- and “can-opener” perforations (producing bony nominates and scapulae are present in inverse flaps), often in places convenient for obtaining a proportion to the relative abundance of hind- and good grip, such as orbits (Fig. 5). They are forelimb elements in the sample. Similarly, there most often found in the orbits, palate, sphenoids, are fewer mandibles in the sample than anticipated maxillae, and parietals. Diameters of punctures from the relatively good representation of crania and “can-opener” perforations range from and partial crania. 2–10 mm, and diameters of nicks from 2–4 mm; Density-mediated attrition, a potentially impor- the majority of these lacerations have a greatest tant factor in assemblage accumulation (Lyman, width of 4–6 mm. 1994c), is unlikely to be responsible for bone In contrast, juvenile monkey crania are usually survivability and fragmentation patterns at found torn open basicranially and missing their Ngogo. Bone-mineral density values for similar- facial portions (Figs. 4 and 6). Infant crania also sized primates and bovids are close enough to one typically lack faces, and are disarticulated along another that the same taphonomic processes their sutures (Figs. 4 and 7a,b). Two associated would not be expected to produce very different parietals from an infant redtail monkey show the bone survivorship patterns (Pickering and effects of the tremendous pressure exerted on the Carlson, 2002). While the patterns of survivorship cranium by a crowned hawk-eagle; one also exhib- in primates and Cephalophus are similar in the its a “can-opener” perforation made by a talon eagle assemblages, the elements that exhibit (Fig. 7b). highest survivability are not those that would be There are also distinctive patterns of postcranial expected based on bone-mineral density values damage in the Ngogo crowned hawk-eagle kill (Pickering and Carlson, 2002). sample (Figs. 2–4). There is a clear tendency in the sample for postcranial bones to be minimally Damage patterns broken or unscathed. For example, a high percentage of long bones exhibit no damage, or have had Despite their great strength and impressive kill- damage limited to the proximal and/or distal ends ing apparatus, crowned hawk-eagles are fastidious of the diaphysis (Figs. 8 and 9a). This is true for all eaters that inflict little damage to bone, compared size classes of long bones (Fig. 9b). Damage to with mammalian predators (Simons, 1966; Mills long bones indicates that crowned hawk-eagles and Mills, 1977; Brain, 1981). Of complete crania remove the ends by repeated slicing with their W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 93

Fig. 2. Bone survivability (percent survivorship, top panels) and fragmentation (percent whole bones, bottom panels) for each primate taxon, as well as the bovid Cephalophus monticola, from the Ngogo crowned hawk-eagle assemblage. 94 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105

Fig. 3. Bone survivability (percent survivorship, top panels) and fragmentation (percent whole bones, bottom panels) for all primates (left) and all non-primate mammals (right) in the Ngogo crowned crowned hawk-eagle assemblage. W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 95

Fig. 4. Bone survivability (percent survivorship, top panels) and fragmentation (percent whole bones, bottom panels) for all preadult (left) and adult (right) primates from the Ngogo crowned hawk-eagle assemblage. 96 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105

Fig. 5. Cercopithecus ascanius cranium from the Ngogo crowned hawk-eagle kill sample. (A) Oblique frontal view. Note talon punctures in orbit and area of infraorbital foramen (arrows). (B) Left lateral view (not to same scale as frontal view). Note perforation in left parietal (arrow) and surrounding bony ﬂaps.

beaks along a narrow circumference. These per- fusion suggest that crowned hawk-eagles find it centages and the evident concentration of effort to difficult to shatter diaphyses to extract marrow. remove ends of long bones near areas of epiphyseal This is supported by the observation that the W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 97

Fig. 6. Inferior view, juvenile cercopithecoid crania from the Ngogo crowned hawk-eagle kill sample. The basicranium and face of each have been removed by crowned hawk-eagles. more robust hindlimb bones are undamaged more 1999; Mitani et al., 2001; Vasquez and Heymann, frequently (66–76%) than the forelimb elements 2001). One of the most formidable of these is (54–68%), and have lower incidences of midshaft the crowned hawk-eagle, which can kill and lift fracture (hindlimb bones, 11–16% vs. forelimb animals considerably heavier than themselves bones, 17–23%) (Fig. 8). (Daneel, 1979; Steyn, 1983). In forest habitats Most of the remaining postcrania are similarly where cercopithecoids are common, they are largely intact, though they may exhibit localized the predominant prey of crowned hawk-eagles fracturing and “can-opener” perforations from (Brown, 1971; Skorupa, 1989; Struhsaker and manipulation by talons (Fig. 10). A notable excep- Leakey, 1990; Mitani et al., 2001). At Ngogo, tion is the severe damage inflicted on scapulae crowned hawk-eagle predation is sufficiently (Fig. 11). The blades of these bones are sharply specialized on redtail monkeys that it has had a shattered and raked, a consequence of the process significant negative impact on their local popula- of opening the body cavity to extract the heart and tion (Mitani et al., 2001). The correlation between lungs (Brown, 1971; Andrews 1990). local availability of species and prey selectivity by these eagles helps explain the difference in selection bias at Ngogo for redtail monkeys and at nearby Discussion Kanyawara in Kibale National Park for colobus monkeys (Brown, 1971; Skorupa, 1989; Struhsaker Raptors are a major class of predators of an- and Leakey, 1990; Mitani et al., 2001). It also thropoid primates (Fowler and Cope, 1964; accounts for contrasting preferences for monkeys Brown, 1971; Rettig, 1978; Seyfarth et al., 1980; and small antelopes in forests and for hyraxes in Gautier-Hion, 1983; Izor, 1985; Eason, 1989; more open-country settings (Snelling and Barbour, Skorupa, 1989; Andrews, 1990; Gargett, 1990; 1969; Brown, 1971; Steyn, 1973, Tuer and Tuer, Heymann, 1990; Peres, 1990; Struhsaker and 1974; Jarvis et al., 1980; Brown, 1982; Brown Leakey, 1990; Sherman, 1991; Cruz-Uribe and et al., 1982; Steyn, 1983; Tarboton and Allen, Klein, 1998; Zhang et al., 1998; Zinner and Pelaéz, 1984; Cruz-Uribe and Klein, 1998). Where prey 98 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105

Fig. 7. Cranial bones of infant cercopithecoids dismembered by crowned hawk-eagles at Ngogo. (A) Parietals and frontals. (B) Left and right parietal bones of a guenon. Note the patent sutures, cracking due to compression, and the “can-opener” talon puncture (arrow). accumulations have been quantiﬁed, it appears Although it is possible that feeding behavior that crowned hawk-eagles narrowly concentrate and ability to fully digest small bones obscures the on one or several prey species (Tuer and Tuer, contribution of small animals to their kill aggrega- 1974; Jarvis et al., 1980; Brown, 1982; Brown et al., tions (Jarvis et al., 1980; Brown, 1982), the recov- 1982; Steyn, 1983; Tarboton and Allen, 1984; ery from below Nests A and B of tiny skeletal Skorupa, 1989; Struhsaker and Leakey, 1990; elements from a number of small mammals Cruz-Uribe and Klein, 1998; Mitani et al., 2001). suggests that the range of taxa in the Ngogo kill W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 99

“can-opener” perforations, but are usually other- wise complete and accompanied by their mandibles. Juvenile crania have their faces and basicrania removed; infant and neonate crania also have their faces broken away, but are usually disarticulated along their sutures and are represented by whole bones (e.g., parietals, occipital). This is surprising given the power of crowned hawk-eagle talons and the fragility of infant and neonate cercopithecoid cranial bones. If the type and frequency of damage to individual bones and bone survivability patterns in the Ngogo sample prove consistent with observations on samples from other sites, they will provide a reliable basis for recognizing the effects of crowned hawk-eagle Fig. 8. Incidence of damage to long bones in the Ngogo predation in fossil assemblages. crowned hawk-eagle kill sample. Tallies are recorded of bones with no damage (“complete”), with damage only to the proxi- To date, the most direct and provocative paleo- mal (“proximal”) or distal (“distal”) end, with damage to both anthropological implication of crowned hawk- proximal and distal ends (“prox. + dist.”), and of bones with eagle predation is Berger and Clarke’s (1995) mid-shaft breakage (“shaft”). hypothesis that this or a similar large raptor may have been primarily responsible for the accumulation of fauna at the late Pliocene site of Taung, sample is adequately represented. By comparison, South Africa, including the infant type specimen of the eclectic prey lists from kill accumulations of Australopithecus africanus. Results of our study crowned hawk-eagles compiled by Jarvis et al. generally support an interpretation of some contri- (1980) are partly artifacts of being composites of bution by large raptors to the Taung fossil assem- observations from many nests; even so, 77% of the blage. For instance, as in the Ngogo crowned largest composite sample is composed of just three hawk-eagle kill sample, the Taung fauna contains taxa (hyrax and two species of bovid). In this a preponderance (85%) of small-to-medium-sized regard, crowned hawk-eagles differ from other animals; it includes a good representation of raptors (e.g., fish eagles, owls), which show a wider cercopithecoid (small baboon) fossil bones and prey selectivity (see, for example, Andrews, 1990; crania; it preserves numerous complete or nearly Stewart et al., 1997, 1999). complete mammalian crania in association with The bone survivability profile of the kill sample mandibles, including that of the Taung child at the of the Ngogo crowned hawk-eagles is also distinc- time of deposition; many of the long bones from tive. Tendencies to dismember and cache animals the Taung fauna are largely complete; and punc- away from the nest and to feed on particular parts ture marks on skulls from Taung exhibit V-shaped of the carcass at the nest contribute to low relative nicks, “can-opener” flaps, and compression cracks, abundances of individual elements. Crania, hind- but not the more severe damage characteristic of limbs, and scapulae are the most commonly repre- mammalian predators (see Brain, 1981). sented elements in bone accumulations below the There are, however, clear differences between nests of crowned hawk-eagles at Ngogo and the kill accumulation from Ngogo and the Taung apparently elsewhere (cf. Brown, 1982). There are sample. The ample presence of hyrax fossils and also clear patterns of bone damage in these aggre- the predominance of baboons suggests that the gations. Long bones tend to be complete or delib- Taung fauna derived from more open conditions erately opened at their ends. Scapular blades are than the prey aggregation from Ngogo, but this heavily damaged and raked. Adult (and subadult) does not preclude eagle involvement. More impor- crania show numerous nicks, punctures, and tantly, the upper size range of mammalian taxa at 100 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 101

Taung, which includes leopards and buffaloes lated, the cranium of the Taung child is much (Cook, 1990; McKee, 1993a,b; Berger and Clarke, larger and more thickly-walled, and therefore 1995), probably exceeds the hunting capabilities of would have been more capable of resisting such eagles, indicating that other agencies of accumula- disintegration, in the manner of a skull from an tion were responsible as well. At the other end of adult cercopithecoid. In addition to its complete- the size spectrum, the association of diverse small ness, the association of a cranium of that size with mammals, eggshells, and crab carapaces at Taung its mandible seems more typical of eagle prey suggested to Brain (1981) the activity of other remains than those of mammalian predators. birds, such as eagle owls. Finally, the puncture and surrounding depressed The Taung fauna is also mixed in another way. fracture in the cranium of the Taung child, pre- While considered by Cooke (1990: p. 122; but see served now as features on the endocast, are con- McKee, 1993a) to represent “a coherent whole... sistent with the type of damage observed on bone essentially from a single fairly short phase of in the Ngogo sample. Thus, while the Taung fauna deposition,” it is actually a composite of fossil appears to represent a mix of deposits and the material from several localities (e.g., “Dart de- behavior of multiple predators, the involvement of posits” and “Hrdlicka deposits”; McKee, 1994), a large raptor like the crowned hawk-eagle in the making it even less likely to have been amassed by death of the Taung child cannot be ruled out. a single animal or type of animal. Indeed, most of Because Berger and Clarke (1995) did not quantify what is known about the overall character of the skeletal damage and bone survivability patterns, Taung fossil assemblage is based on taxa from the however, the taphonomic impact of a predator Hrdlicka deposits and not from the more faunally such as the crowned hawk-eagle in the Taung depauperate hominid type site (McKee, 1994). faunal assemblage remains unclear. Further assess- The likelihood of faunal mixing and multiple ment of their hypothesis will require a more com- accumulating agents makes it difficult to evaluate prehensive examination of the faunal collection rigorously the possibility of eagle involvement in from the hominid type site at Taung. the death of the Taung child. Furthermore, there is no evidence of crowned hawk-eagles or other raptors attempting to take infant chimpanzees, Summary which are similar in size to the estimated body mass of 10–12 kg for the Taung child (Berger and Taphonomic analysis of a prey assemblage of Clarke, 1995), at Ngogo or elsewhere (Mitani crowned hawk-eagles from the Ngogo study area et al., 2001). Nevertheless, a crowned hawk-eagle in Kibale National Park, Uganda revealed distinct was observed to attack a human child, and in one patterns of prey selection, bone damage, and sur- instance an infant human cranium was discovered vivability of skeletal elements. The animals in the in a crowned hawk-eagle nest (Steyn, 1983). prey assemblage are small-to-medium-sized, rang- Although no young primate killed by crowned ing up to about 11 kg in body mass, and are hawk-eagles at Ngogo is larger than a juvenile predominantly cercopithecoid monkeys (81%). baboon, these cases and the size range of crowned Redtail monkeys are particularly well represented hawk-eagle prey suggest that it is reasonable to in this faunal accumulation, composing 66% of the believe that they were capable of killing an infant cercopithecoid monkeys identifiable to species. australopithecine. Despite the fact that infant pri- This strong prey selectivity may reflect the contri- mate skulls in the Ngogo sample are all disarticu- bution that monkeys make to the mammalian

Fig. 9. (A) Cercopithecoid long bones from the Ngogo crowned hawk-eagle kill sample with proximal and distal ends removed for the extraction of marrow. (B) Cercopithecoid femora from the Ngogo crowned hawk-eagle kill sample. From left to right, two neonate femora, and adult femora of Cercopithecus ascanius (female and male individuals), Cercopithecus lhoesti, Colobus guereza, Piliocolobus badius,andLophocebus albigena. 102 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105

Fig. 10. Cercopithecoid innominates from the Ngogo crowned hawk-eagle kill sample. (A) Left and right innominates, Cercopithecus ascanius. Note the talon damage to the ilia and pubic areas. (B) Enlarged detail of a “can-opener” perforation (arrow) through the ilium of the right innominate (#151), lateral view. W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 103

Fig. 11. Cercopithecus ascanius scapulae from the Ngogo crowned hawk-eagle kill sample. Note the damage caused by raking and perforation with talons. biomass in the Ngogo area, and the local avail- taphonomic signature of such large raptors in ability of species. Conversely, the Ngogo kill fossil assemblages. Indeed, comparison of infor- sample shows little or no sign of preference for mation on the fossil fauna from the late Pliocene animals of a particular sex or age-class. site of Taung, South Africa with the results of our In addition to an evident concentration on analysis indicates that crowned hawk-eagles or a particular taxa, the Ngogo kill sample also exhibits similar large raptor had a role in the accumulation conspicuous features of bone survival and damage, of that fauna, including the infant type skull of including: (1) better representation of crania, hind- Australopithecus africanus. Future study of the limb elements, and scapulae than other elements; Taung fauna in a manner similar to our analysis of (2) tendency for adult crania to remain complete the Ngogo kill sample is necessary for a more and to be accompanied by their mandibles; (3) rigorous evaluation of the extent of contribution of faces and basicrania are destroyed, and mandibles large raptors to the fossil deposit from the hominid dissociated from crania in young individuals; (4) type site and the probability that such a predator mandibles are largely undamaged; (5) long bones killed the Taung child. tend to be complete or opened only at their ends; (6) thin bones, such as crania and innominates, exhibit numerous nicks, punctures, and “can- opener” perforations; and (7) scapular blades are Acknowledgements heavily damaged and raked. If these features are consistent for crowned hawk-eagle kill assem- Research at Ngogo was sponsored by the blages, they may prove useful for recognizing the Makerere University, the Ugandan National 104 W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105

Parks, and the Ugandan Wildlife Authority. We angels: essays in anthropology in honor of Phillip V. thank John Kasenene, Gilbert Isabirye-Basuta, Tobias. Wiley-Liss, New York, pp. 119–134. and the staff of the Makerere University Biological Cordeiro, N.J., 1992. Behaviour of blue monkeys (Cerco- pithecus mitis) in the presence of crowned eagles (Stephanoa- Field Station for providing logistical assistance. etus coronatus). Folia Primatol. 59, 203–207. We are grateful to Charles Businge, Adolf Cruz-Uribe, K., Klein, R.G., 1998. Hyrax and hare bones from Magoba, Godfrey Mbabazi, Lawrence Ndangizi, modern South African eagle roosts and the detection of Tom Struhsaker, Alfred Tumusiime, David Watts, eagle involvement in fossil bone assemblages. J. Archaeol. Jeremiah Lwanga, and Tammy Windfelder for Sci. 25, 135–147. Daneel, A.B.C., 1979. Prey size and hunting methods of the help in the field. Bonnie Miljour skillfully assisted crowned eagle. Ostrich 50, 120–121. with preparation of the figures. Catherine Badgley Delson, E., Terranova, C.J., Jungers, W.L., Sargis, E.J., helped with understanding methods of calculating Jablonski, N.G., Dechow, P.C., 2000. Body mass in Cerco- minimum numbers of individuals. Eric Delson pithecidae (Primates, Mammalia): estimation and scaling in generously shared data and gave access to primate extinct and extant taxa. Am. Mus. Nat. Hist. Anthropol. Papers 83, 1–159. skeletal material in his care. Permission to study Eason, P., 1989. Harpy eagle attempts predation on adult primate skeletal collections at the American howler monkey. Condor 91, 469–470. Museum of Natural History, New York City was Fowler, J.M., Cope, J.B., 1964. Notes on the harpy eagle in granted by Ross MacPhee. Loan of a crowned British Guiana. Auk 81, 257–273. hawk-eagle skeleton from the National Museum of Gargett, V., 1990. The black eagle, a study. John Voelcker Bird Natural History, Washington, D.C. was expedited Book Fund, Johannesburg. Gautier-Hion, A., 1975. Dimorphisme sexuel et organisation by James Dean. We much appreciate the careful sociale chez les cercopithecines forestiers africains. review of the paper and many helpful suggestions Mammalia 39, 365–374. to improve it provided by Peter Andrews, Lee Gautier-Hion, A., Tutin, C.E.G., 1988. Simultaneous attack by Lyman, and Terry Harrison. This research was adult males of a polyspecific troop of monkeys against a supported by an NSF Presidential Faculty Fellow- crowned hawk eagle. Folia Primatol. 51, 149–151. Gautier-Hion, A., Quris, R., Gautier, J.P., 1983. Monospecific ship Award SBR-9253590 and grants from the vs. polyspecific life: a comparative study of foraging and National Geographic Society and Wenner-Gren antipredator tactics in a community of Cercopithecus Foundation for Anthropological Research to monkeys. Behav. Ecol. Sociobiol. 12, 325–335. JCM, and funds from the Museum of Paleontol- Grayson, D.K., 1984. Quantitative zooarchaeology. Academic ogy, University of Michigan to W.J.S. Press, Orlando. Hart, J.A., Katembo, M., Punga, K., 1996. Diet, prey selection and ecological relations of leopard and golden cat in the Ituri Forest. Zaire Afr. J. Ecol. 34, 364–379. Heymann, E.W., 1990. Reactions of wild tamarins Saguinus References mystax and Saguinus fuscicollis to avian predators. Int. J. Primatol. 11, 327–337. Andrews, P., 1990. Owls, caves and fossils. University of Izor, R., 1985. Sloths and other mammalian prey of the harpy Chicago Press, Chicago. eagle, in: Montgomery, G. (Ed.), The evolution and ecology Badgley, C., 1986. Counting individuals in mammalian fossil of armadillos, sloths and vermilinguas. Smithsonian assemblages from fluvial environments. Palaios 1, 328–338. Institution Press, Washington, D.C., pp. 343–346. Berger, L.R., Clarke, R.J., 1995. Eagle involvement in accumu- Jarvis, M.J.F., Currie, M.H., Palmer, N.G., 1980. Food of lation of the Taung child fauna. J. Hum. Evol. 29, 275–299. crowned eagles in the Cape Province, South Africa. Ostrich Brain, C.K., 1981. The hunters or the hunted? University of 51, 215–218. Chicago Press Chicago. Jones, W.T., Bush, B., 1980. Darting and marking techniques Brown, L.H., 1971. The relations of the crowned eagle Steph- for an arboreal forest monkey, Cercopithecus ascanius. Am. anoaetus coronatus and some of its prey animals. Ibis 113, J. Primatol. 14, 83–89. 240–243. Kingdon, J., 1997. The Kingdon field guide to African Brown, L.H., 1982. The prey of the crowned eagle Stephanoa- mammals. Academic Press, London. etus coronatus in Central Kenya. Scopus 6, 91–94. Klein, R.G., Cruz-Uribe, K., 1984. The analysis of animal Brown L., Urban E., Newman K., 1982. The birds of Africa, bones from archeological sites. University of Chicago Press, vol. 1. Academic Press, London. Chicago. Cooke, H.B.S., 1990. Taung fossils in the University of Leland, L., Struhsaker, T.T., 1987. Monkey business. Anim. California collections, in: Sperber, G.H. (Ed.), From apes to Kingdom 90, 24–37. W.J. Sanders et al. / Journal of Human Evolution 44 (2003) 87–105 105

Leland, L., Struhsaker, T.T., 1993. Teamwork tactics. Nat. Rettig, N.L., 1978. Breeding behavior of the harpy eagle Hist. 102, 43–48. (Harpia harpyja). Auk 95, 629–643. Lyman, R.L., 1994a. Quantitative units and terminology in Seyfarth, R.M., Cheney, D.L., Marler, P., 1980. Monkey zooarchaeology. Am. Antiq. 59, 36–71. responses to three different alarm calls: evidence of predator Lyman, R.L., 1994b. Relative abundances of skeletal specimens classification and semantic communication. Science 210, and taphonomic analysis of vertebrate remains. Palaios 9, 801–804. 288–298. Sherman, P.T., 1991. Harpy eagle predation on a red howler Lyman, R.L., 1994c. Vertebrate taphonomy. Cambridge monkey. Folia Primatol. 56, 53–56. University Press, Cambridge. Simons, J.W., 1966. The presence of leopard and a study of the Maisels, F.G., Gautier, J.-P., Cruickshank, A., Bosefe, J.-P., food debris in the leopard lairs of the Mount Suswa caves, 1993. Attacks by crowned hawk eagles (Stephanoaetus coro- Kenya. Bull. Cave Explor. Gp. E. Afr. 1, 51–69. natus) on monkeys in Zaire. Folia Primatol. 61, 157–159. Skorupa, J.P., 1989. Crowned eagles Stephanoaetus coronatus Marean, C.W., Abe, Y., Nilssen, P.J., Stone, E.C., 2001. in rainforest: observations on breeding chronology and diet Estimating the minimum number of skeletal elements at a nest in Uganda. Ibis 131, 294–298. (MNE) in zooarchaeology: a review and a new image- Smith, R.J., Jungers, W.L., 1997. Body mass in comparative analysis GIS approach. Am. Antiq. 66, 333–348. primatology. J. Hum. Evol. 32, 523–559. McKee, J.K., 1993a. Faunal dating of the Taung hominid fossil Snelling, J.C., Barbour, D.Y., 1969. Crowned eagle. deposit. J. Hum. Evol. 25, 363–376. Bokmakierie 21, 67–69. Stewart, K.M., Matthiesen, D.P., Leblanc, L., West, J., 1997. McKee, J.K., 1993b. Formation and geomorphology of caves Prey diversity and selectivity of the African fish eagle: data in calcareous tufas and implications for the study of the from a roost in northern Kenya. Afr. J. Ecol. 35, 133–145. Taung fossil deposits. Trans. R. Soc. S. Afr. 48, 307–322. Stewart, K.M., Leblanc, L., Matthiesen, D.P., West, J., 1999. McKee, J.K., 1994. Catalogue of fossil sites at the Buxton Microfaunal remains from a modern east African raptor Limeworks. Taung. Pal. Afr. 31, 73–81. roost: patterning and implications for fossil bone scatters. Mills, M.G.L., Mills, M.E.J., 1977. An analysis of bones Paleobiology 25, 483–503. collected at hyaena breeding dens in the Gemsbok National Steyn, P., 1973. Eagle days. A study of African eagles at the Parks (Mammalia: Carnivora). Ann. Transvaal. Mus. 30, nest. Purnell, Johannesburg. 146–154. Steyn, P., 1983. Birds of prey of Southern Africa. Tanager, Mitani, J.C., Watts, D., 2001. Why do chimpanzees hunt and Dover. share meat? Anim. Behav. 61, 915–924. Struhsaker, T.T., 1997. Ecology of an African rain forest. Mitani, J.C., Struhsaker, T.T., Lwanga, J.S., 2000. Primate University Press of Florida, Gainesville. community dynamics in old growth forest over 23.5 years at Struhsaker, T.T., Leakey, M., 1990. Prey selectivity by crowned Ngogo, Kibale National Park, Uganda: implications for hawk-eagles on monkeys in the Kibale Forest, Uganda. conservation and census methods. Int. J. Primatol. 21, Behav. Ecol. Sociobiol. 26, 435–443. 269–286. Strum, S.C., 1991. Weight and age in wild olive baboons. Am. Mitani, J.C., Sanders, W.J., Lwanga, J.S., Windfelder, T.L., J. Primatol. 25, 219–237. 2001. Predatory behavior of crowned hawk-eagles Tarboton, W.R., Allen, D.G., 1984. The status and conserva- (Stephanoaetus coronatus) in Kibale National Park, tion of birds of prey in the Transvaal, Transvaal Museum Uganda. Behav. Ecol. Sociobiol. 49, 187–195. Monograph. vol. 3. Transvaal Museum, Pretoria. Nakatsukasa, M., 1994a. Intrageneric variation of limb bones Tuer, V., Tuer, J., 1974. Crowned eagles of the Matopos. and implications for positional behavior in Old World Honeyguide 80, 33–41. monkeys. Z. Morphol. Anthrop. 80, 125–136. Vasquez, M.R.O., Heymann, E.W., 2001. Crested eagle (Mor- Nakatsukasa, M., 1994b. Morphology of the humerus and phnus guianensis) predation on infant tamarins (Saguinus femur in African mangabeys and guenons: functional adap- mystax and Saguinus fuscicollis, Callitrichinae). Folia tation and implications for the evolution of positional Primatol. 72, 301–303. behavior. Afr. Study. Monogr. Suppl. 21, 1–61. Williams, J.G., Arlott, N., 1980. The Collins Field Guide to the Nakatsukasa, M., 1996. Locomotor differentiation and differ- birds of East Africa. Stephen Greene, Lexington. ent skeletal morphologies in mangabeys (Lophocebus and Zhang, S., Ren, B., Li, B., 1998. A juvenile Sichuan golden Cercocebus). Folia Primatol. 66, 15–24. monkey (Rhinopithecus roxellana) predated by a goshawk Peres, C.A., 1990. A harpy eagle successfully captures an adult (Accipiter gentilis) in the Qinling Mountains. Folia howler monkey. Wilson Bull. 102, 560–561. Primatol. 70, 175–176. Pickering, T.R., Carlson, K.J., 2002. Baboon bone mineral Zinner, D., Pelaéz, F., 1999. Verreaux’s eagles (Aquila ver- densities: implications for the taphonomy of primate reauxi) as potential predators of hamadryas baboons (Papio skeletons in South African cave sites. J. Arch. Sci. 29, hamadryas hamadryas) in Eritrea. Am. J. Primatol. 47, 883–896. 61–66.